Lach

1. std.foldl doesn't use recursion Your code, however, does, due to lazy evaluation 2. os-stack does nothing here, only max-stack will help max-stack specifies how much recursion is allowed, os-stack...

Making this call tailstrict `for field in std.objectFields(manifest.spec.template.nodeSelectors) tailstrict` greately reduces stack size usage, due to field nodeSelectors being calculated prior to getting it fields, instead of going 500 overlays...

Oh, so it happens in the real code, not this one. Yes, only os-stack will help here then. Unfortunately, Rust/C++ are unable to recover from stack overflow, only golang is...

I have implemented dynamic stack growth on master, so you should not experience need to use --os-stack anymore. However, max-stack is still there, and its limit is not changed. I...

While this is not a permanent solution, this issue is resolved for now, with better solution in development.

Is it? ``` out ├── a.jsonnet ├── b.jsonnet └── library └── c.jsonnet ``` a.jsonnet: ```jsonnet import 'b.jsonnet' ``` b.jsonnet: ```jsonnet import 'c.jsonnet' ``` c.jsonnet: ```jsonnet 1 + 2 ``` Running...

Don't you have symlinks lying around, by any chance?

For long-running applications you need to collect garbage sometimes, it is not fully automatic (There was a branch with automated GC based on linux's memory pressure information, but it is...

Huh. I wonder why jpath is not resolved in input in official jsonnet, yet tla-code-file uses it.

You shouldn't need this patch on 6.16 kernel, since VivePro2 announces its own usage using EDID, and kernel is able to detect that automatically